I dont think replacing those artic cooler fans is going to do anything for preventing those caps or VRMs from burning out. Maybe adding or reorienting case fans may help if those components are overheating (are they?)
HUGE DISCLAIMER I'm going by pictures posted not my own eyeballs looking at the PCB's personally!
I am just guessing at a possibility, let me repeat, I am just guessing at a possibility.
(^^ this will not prevent a tangent but I tried ^^)
The latest pics show the explosion seems like it happened on a secton of bare PCB right next to a chipcap.
The component looks intact.
How could this happen?
A FEW of the possibilities are:
Loose metal on PCB.
Poor PCB etching. (why would it wait to explode if so?)
Metal contamination could also account for performance issues and failure seen by others.
The explosions might only be happening at locations where there is high current.
The performance issues might be happening in low current locations.
Where could the metal come from?
A problem with the solder paste or application process.
A solder reflow problem. (cold solder joints)
Metal dirt on case leftover from manufacture.
Metal dirt on heatsinks leftover from manufacture.
Self tapping case screws or standoffs might create metal shavings.
A serious mechanical problem with a machine that makes the PCB's.
I DO NOT RECOMMEND ANYONE OPEN THE CASE JUST TO LOOK
But if someone has a unit and already has the case open:
They could use very bright light and a magnifying glass to look.
Move the light to view with different light angles and shadows.
Look at the heatsink and edges of the case for metal shavings.
Up until I saw the last pics I was thinking faulty component issue or a programming quirk overloading good components. Now I'm not so sure. A spark away from components is usually caused by foreign objects.
Did I mention that I am just guessing?
I think I did mention that I am just guessing at a possibility at least once.